NPTEL Introduction To Soft Computing Week 8 Assignment Answers 2025
1. Which of the following statements best describes the learning process in Artificial Neural
Networks (ANNs)?
a. Learning in ANNs involves updating weights, biases, and other parameters through training.
b. Learning in ANNs is a temporary process and does not lead to long-term changes in behavior:
c. Learning in ANNs is purely based on predefined rules and does not require experience.
d. Learning in ANNs does not involve optimization and is independent of curve fitting.
Answer :- For Answers Click Here
2. Which of the following statements is TRUE regarding learning techniques in neural networks?
a. In supervised learning, the network learns without any target output or error feedback.
b. Unsupervised learning requires a labeled dataset for training.
c. Supervised learning involves updating the network using error feedback based on known input-output pairs.
d. Unsupervised learning compares the network’s output with a predefined target value to minimize error.
Answer :-
3. Which of the following statements best describes reinforcement learning?
a. The teacher provides the exact expected output for every input.
b. The network receives feedback in the form of rewards and penalties rather than explicit target outputs.
c. Reinforcement learning is identical to supervised learning since it requires labeled data.
d. It is not useful in artificial neural networks due to the lack of training data.
Answer :-
4. What is the primary objective of the gradient descent algorithm in neural networks?
a. To maximize the error function for better generalization.
b. To iteratively adjust network parameters to minimize the loss function.
c. To randomly update weights and biases without any optimization strategy.
d. To ensure the neural network never overfits the training data.
Answer :-
5. What is the primary characteristic of competitive learning in neural networks?
a. All neurons update their weights simultaneously based on input stimuli.
b. Neurons that respond most strongly to the input have their weights updated, following a Winner-Takes-All strategy.
c. The network learns by minimizing an error function similar to supervised learning.
d. All neurons equally share weight updates to ensure uniform learning across the network.
Answer :-
6. What is the role of the learning rate in training a neural network?
a. It controls the step size for updating the weights during training.
b. It determines the number of epochs in the training process.
c. It ensures that the model always converges to the global minimum.
d. It eliminates the need for backpropagation in deep learning models.
Answer :- For Answers Click Here
7. Which of the following best describes the purpose of the Backpropagation Algorithm in
neural networks?
a. It helps in updating weights by propagating errors from the output layer back to the input layer.
b. It selects the best activation function for the network.
c. It increases the number of neurons in the hidden layers dynamically.
d. It eliminates the need for training data by learning in an unsupervised manner.
Answer :-
8.
Answer :-
9. Which of the following statements about the MATLAB Fuzzy Toolbox is TRUE?
a. The MATLAB Fuzzy Toolbox consists of the FIS Editor and the Fuzzy Controller.
b. The FIS Editor is used to define and modify Fuzzy Inference System (FIS) variables.
c. The Fuzzy Controller is a Simulink block that implements the FIS rules.
d. All of the above.
Answer :-
10. How can you start the MATLAB ANN (Artificial Neural Network) Toolbox?
a. Type fuzzy in the MATLAB command line.
b. Type instart in the MATLAB command line.
c. Type anfis in the MATIAB command line.
d. Type simulink in the MATLAB command line.
Answer :- For Answers Click Here
